SAN FRANCISCO (AP) — A study of how three popular artificial intelligence chatbots respond to queries about suicide found that they generally avoid answering questions that pose the highest risk to the user, such as for specific how-to guidance.
Read more >